An assessment of the tropical Humidity Temperature covariance - - PowerPoint PPT Presentation

An assessment of the tropical Humidity – Temperature covariance using AIRS

Antonia Gambacorta, UMBC/PSGS Chris Barnet, NOAA/NESDIS Brian Soden, Univ. of Miami Larrabee Strow, UMBC

AIRS SCIENCE MEETING, October 10, 2007

2

Outline

• We investigate the horizontal and vertical structure of covariance between

water vapor and temperature in the tropical troposphere using AIRS

• We compare with previous study (radiosondes, ECMWF, GCMs) which

have focused only on the covariance of tropical mean quantities, and have shown a general uniform positive correlation through the whole troposphere

• AIRS high spectral resolution and the radiance cloud clearing algorithm

allow for high vertical resolution and excellent spatial coverage

• respectively. This enables a more comprehensive analysis than has

previously been possible.

• AIRS shows large spatial gradient in the local covariance between water

vapor and temperature – Submitted paper: A. Gambacorta, C. Barnet, B. Soden, L. Strow, “An assessment of the tropical humidity-temperature covariance using AIRS”, GRL, 2007

• Implications for climate

3

Motivations

• Lindzen hypothesis: the vertical distribution of water vapor in

the tropics is characterized by three distinct regions: the convective domain of the boundary layer, the free tropical troposphere and the

• utflow domain of deep convection: the dependence of water vapor on

temperature may be strongly height dependent Water vapor is the most active greenhouse gas in regulating the radiation budget of the atmosphere

• The Clausius –Claperyon equation: es~exp(-1/T)

Temperature Moisture content Positive Feedback

In the Upper Troposphere (UT): In the Upper Troposphere (UT):

Temperature Convective Tower Height Precipitation UT moisture content Drying negative Feedback

4

Fractional change of q wrt T

q d dOLR ln

• m

m a dT

dq q dT q d 1 ~ ln

qa=Annual mean specific humidity qm, Tm= monthly mean specific humidity and temperature

[Ref.: Huang and Soden, GRL, 2005; Sun and Oort, J.Climate,1995;] (Ref.: Huang and Soden, GRL, 2005)

5

AIRS Dataset

• 3x3 degree gridded subsets of AIRS products (no spatial

bias) from August 2003 to April 2007.

• NOAA emulation of version 5 retrieval algorithm, using a

version 4–like rejection criterium:

– To preserve whole accepted retrieval profiles – To increase sampling in tropical cloudy convective regions.

• KEY Elements of AIRS database for this analysis:

– Cloud clearing: increase of the daily yield of observational data up to 80% (no clear-sky bias typical of satellite measurements) – Accurate retrieval algorithm: ~1 K for temperature; ~10 and 20% rms for water vapor in the tropical lower-middle and upper troposphere respectively – Uniform spatial coverage and high vertical resolution of ~2-3km for T and ~ 2km for WV

6

Fractional change of q wrt T, from AIRS [ , %/K ]

300mb 600mb 850mb

Clausius-Clapeyron regime ~7%/K

7

AIRS (left) vs NOAA GFDL model (right)

300mb 600mb 850mb August 2003 – April 2007 January 1998 – December 2004

8

Comparison with the NCEP vertical velocity field

August 2003 – December 2004

9

Fractional change of q wrt T, from AIRS [ , %/K ]

• AIRS shows a strong latitude-longitude dependence in the structure of the

water vapor – temperature covariance, particularly in the free troposphere, where extended negative and positive covariance up to one order of magnitude larger than the Clausius-Claperyon eq. are found

• Highest positive responses found in the upper troposphere region
• Highest negative responses found in the middle troposphere region
• Comparisons with the NOAA GFDL model show same order of magnitude

in the variability of the water vapor – temperature cavariance values

• Other mechanisms regulating the water vapor distribution in the free tropical

troposphere besides local temperature, appear to be connected to the patterns of the large scale circulation: regions of positive and negative covariance roughly resemble regions of the ascending and descending branches, respectively,

• f the tropical circulation

10

Zonal cross section

AIRS GFDL

11

Factional changes of q and T tropical averages

__ AIRS 3x3 degree spatial resolution __ GFDL full spatial resolution __ Const RH hypothesis

• - AIRS -- GFDL on a common spatial subset
• .- AIRS further degraded on same vertical res of GFDL
• When we average over the whole tropical domain, the regression slopes are of the same order
• f magnitude of the Clausius Claperyon regime at ALL levels in the troposphere

(Ref.: Huang and Soden, GRL, 2005)

12

Conclusions

• Exploiting the high vertical resolution and excellent spatial coverage, the AIRS

instrument shows a complex horizontal and vertical structure of the humidity- temperature covariance

• In the upper troposphere region, water vapor appears to be most strongly

and overall positively tied to local temperature changes

• Negative correlations characterize extended regions of the free troposphere,

particularly at mid altitude levels ( ~600mb) where tropically averaged correlations become negative.

• Values up to one order of magnitude larger than the

Clausius-Clapeyron regime, suggests that other processes besides local temperature, play a more important role in determining moisture changes in the free troposphere, and appear to be connected to the transport mechanisms of the large-scale tropical circulation.

13

Future works:

• Recent findings [Vecchi et al., Nature, 2006] of a

weakening process of the tropical circulation due to anthropogenic forcing lead to new questions:

• What are these moistening and drying sources?
• How do they relate to the tropical circulation?
• What is their overall radiative role in the Earth’s

energy budget?

14

15